CN116578942B - Method and device for processing list exception - Google Patents
Method and device for processing list exception Download PDFInfo
- Publication number
- CN116578942B CN116578942B CN202310853781.8A CN202310853781A CN116578942B CN 116578942 B CN116578942 B CN 116578942B CN 202310853781 A CN202310853781 A CN 202310853781A CN 116578942 B CN116578942 B CN 116578942B
- Authority
- CN
- China
- Prior art keywords
- abnormal
- list
- detection
- information
- sample information
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000012545 processing Methods 0.000 title claims abstract description 79
- 238000000034 method Methods 0.000 title claims abstract description 76
- 230000002159 abnormal effect Effects 0.000 claims abstract description 237
- 238000001514 detection method Methods 0.000 claims abstract description 213
- 230000005856 abnormality Effects 0.000 claims abstract description 55
- 230000009123 feedback regulation Effects 0.000 claims abstract description 25
- 238000011156 evaluation Methods 0.000 claims abstract description 17
- 230000001174 ascending effect Effects 0.000 claims description 16
- 238000012549 training Methods 0.000 claims description 13
- 238000007637 random forest analysis Methods 0.000 claims description 6
- 239000013585 weight reducing agent Substances 0.000 claims description 6
- 235000019633 pungent taste Nutrition 0.000 claims description 5
- 238000012550 audit Methods 0.000 abstract description 19
- 230000007246 mechanism Effects 0.000 abstract description 12
- 230000000694 effects Effects 0.000 abstract description 11
- 238000007726 management method Methods 0.000 abstract description 9
- 230000015654 memory Effects 0.000 description 32
- 230000008569 process Effects 0.000 description 14
- 238000004422 calculation algorithm Methods 0.000 description 12
- 238000010586 diagram Methods 0.000 description 11
- 230000006399 behavior Effects 0.000 description 10
- 230000001680 brushing effect Effects 0.000 description 9
- 230000033228 biological regulation Effects 0.000 description 8
- 230000008859 change Effects 0.000 description 5
- 238000000855 fermentation Methods 0.000 description 5
- 230000004151 fermentation Effects 0.000 description 5
- 238000012544 monitoring process Methods 0.000 description 4
- 238000003825 pressing Methods 0.000 description 4
- 238000011160 research Methods 0.000 description 4
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 3
- 208000035475 disorder Diseases 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 238000003672 processing method Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 230000003370 grooming effect Effects 0.000 description 2
- 238000012423 maintenance Methods 0.000 description 2
- 230000004044 response Effects 0.000 description 2
- 238000012552 review Methods 0.000 description 2
- 230000001360 synchronised effect Effects 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- KLDZYURQCUYZBL-UHFFFAOYSA-N 2-[3-[(2-hydroxyphenyl)methylideneamino]propyliminomethyl]phenol Chemical compound OC1=CC=CC=C1C=NCCCN=CC1=CC=CC=C1O KLDZYURQCUYZBL-UHFFFAOYSA-N 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 230000002411 adverse Effects 0.000 description 1
- 238000003491 array Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 201000001098 delayed sleep phase syndrome Diseases 0.000 description 1
- 208000033921 delayed sleep phase type circadian rhythm sleep disease Diseases 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000010921 in-depth analysis Methods 0.000 description 1
- 230000010365 information processing Effects 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000011946 reduction process Methods 0.000 description 1
- 230000000630 rising effect Effects 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 230000001960 triggered effect Effects 0.000 description 1
Landscapes
- Testing And Monitoring For Control Systems (AREA)
Abstract
The embodiment of the application relates to a method and a device for processing list exceptions, wherein the method comprises the following steps: acquiring target list information, and carrying out anomaly detection on the target list information according to a set detection method to obtain a corresponding anomaly detection result; inputting the abnormal sample information corresponding to the abnormal detection result into a pre-trained estimated model for evaluation processing, and outputting the on-board duration corresponding to the abnormal sample information; determining a feedback regulation strategy according to the on-board duration; and executing the processing of the abnormal on-board information based on the feedback adjustment strategy. Detecting abnormal sample information in each piece of list information by creating a detection tool for list abnormality, and processing the abnormal sample information by a set feedback adjustment strategy to achieve the purpose of managing the abnormal list information; therefore, the technical effects of forming a set of hot list management mechanism for real-time alarming, feedback and adjustment and maintaining fairness and stability of the hot list by combining machine audit with manual audit can be achieved.
Description
Technical Field
The embodiment of the application relates to the technical field of list information processing, in particular to a method and a device for processing list abnormality.
Background
In the age of information overload, the recommendation algorithm comprises a personalized recommendation algorithm and a personalized recommendation system, the recommendation algorithm also comprises non-personalized recommendation, and the typical application scene is various types of sheets. Non-personalized recommendation is applied more in the early stage of the Internet age, but as search engines rise, personalized recommendation scenes are increased, and the non-personalized recommendation scenes are gradually compressed.
Currently, non-personalized recommendation algorithms are generally classified into time sensitive algorithms, heat sensitive algorithms, or a mixture of both. With the gradual increase of the influence of the hot search list or the ranking list on the upper list information, some organizations can attack the list in the interest, influence the fairness of the ranking algorithm and the list, and cause algorithm disorder of the list abnormality. The 'list up and down list abnormality' as one of the 'list abnormality' can cause the loss of list fairness and can influence the public confidence of the list and influence the platform benefit. In addition, too many top-level game content with business purposes and entertainment can influence the user experience and bring adverse effects to the network environment.
At present, researches related to 'list abnormality up and down' in sorting problems are few, most of the researches are only performed on a certain topic to track the list, or statistics and discovery are performed on intervention possibly existing in the list, and most of the researches are performed on certain contents in the list, such as certain topics and social robots, in-depth analysis is performed, so that researches on the list are stopped in the aspects of statistical analysis, list validity verification and the like.
Disclosure of Invention
In view of this, in order to solve the above-mentioned technical problem of lack of detection and treatment of the list abnormality, the embodiments of the present application provide a method and apparatus for processing a list abnormality.
In a first aspect, an embodiment of the present application provides a method for processing a list exception, including:
acquiring target list information, and performing anomaly detection on the target list information according to a set detection method to obtain a corresponding anomaly detection result;
inputting the abnormal sample information corresponding to the abnormal detection result into a pre-trained pre-estimated model for evaluation processing, and outputting the on-board duration corresponding to the abnormal sample information;
determining a feedback adjustment strategy according to the on-board duration;
and executing the processing of the abnormal list information based on the feedback adjustment strategy.
In one possible implementation manner, the detecting the abnormality of the target list information according to the set detection method includes:
performing anomaly detection on the target list information according to a set detection rule;
and/or the number of the groups of groups,
and carrying out anomaly detection on the target list information according to a set detection model.
In one possible implementation manner, the performing anomaly detection on the target list information according to a set detection rule includes:
Detecting the variation range of the target list information in the list position according to a set first detection rule;
and, a step of, in the first embodiment,
detecting the same position on-board duration of the target list information according to a set second detection rule;
and, a step of, in the first embodiment,
performing hotness ranking consistency detection on the target list information according to a set third detection rule;
and, a step of, in the first embodiment,
and detecting the abnormality of the main list and the auxiliary list of the target list according to the set fourth detection rule.
In one possible implementation manner, the performing anomaly detection on the target list information according to a set detection model includes:
acquiring target characteristics corresponding to the target list information;
and inputting the target characteristics into a pre-trained random forest classification model to perform anomaly detection.
In one possible implementation manner, the obtaining the corresponding abnormality detection result includes:
when abnormal sample information is detected according to the detection rule, a corresponding rule abnormal result is obtained;
when abnormal sample information is not detected according to the detection rule, a corresponding rule normal result is obtained;
and/or the number of the groups of groups,
when abnormal sample information is detected according to the detection model, a corresponding classified abnormal result is obtained;
And when abnormal sample information is not detected according to the detection model, a corresponding classification normal result is obtained.
In one possible implementation manner, the inputting the abnormal sample information corresponding to the abnormal detection result into a pre-trained pre-estimated model for evaluation, and outputting the on-board duration corresponding to the abnormal sample information includes:
acquiring abnormal sample characteristics of the abnormal sample information;
and inputting the abnormal sample characteristics into a pre-trained pre-estimated model to carry out regression processing of a weighted least square method, so as to obtain the abnormal list duration.
In a possible implementation manner, the determining a feedback adjustment policy according to the on-board duration includes:
determining corresponding normal sample information according to the abnormal sample information;
inputting the abnormal sample information and the normal sample information into the pre-estimated model to determine a time length gain;
determining a feedback adjustment parameter according to the on-board duration of the abnormal sample information and the corresponding duration gain;
when the feedback regulation parameter is larger than a set parameter threshold value, determining an ascending order feedback regulation strategy;
and determining a descending order feedback regulation strategy when the feedback regulation parameter is smaller than a set parameter threshold value.
In one possible implementation manner, the processing the abnormal sample information based on the feedback adjustment strategy includes:
based on an ascending order feedback adjustment strategy, carrying out weighting processing on the on-board position of the abnormal sample information according to the feedback adjustment parameters;
and based on the descending feedback adjustment strategy, performing weight reduction processing on the in-list position of the abnormal sample information according to the feedback adjustment parameters.
In one possible implementation manner, before the target feature corresponding to the target list information is obtained, the method further includes:
acquiring history list information and acquiring history list characteristics corresponding to the history list information;
setting up a rule detection tool based on the detection rule set by the history list information, and setting up a model detection tool based on the history list characteristics;
performing anomaly detection on the history list information based on the model detection tool and the rule detection tool to obtain a history anomaly result;
and inputting the historical abnormal information corresponding to the historical abnormal result into a pre-training model for pre-training to obtain a trained pre-estimation model.
In a second aspect, an embodiment of the present application provides a processing apparatus for list exception, including:
The anomaly detection module is used for acquiring target list information, and carrying out anomaly detection on the target list information according to a set detection method to obtain a corresponding anomaly detection result;
the evaluation module is used for inputting the abnormal sample information corresponding to the abnormal detection result into a pre-trained estimated model for evaluation processing and outputting the on-board duration corresponding to the abnormal sample information;
the determining module is used for determining a feedback adjustment strategy according to the on-board duration;
and the processing module is used for executing the processing of the abnormal list information based on the feedback adjustment strategy.
According to the list abnormality processing scheme provided by the embodiment of the application, the target list information is obtained, and abnormality detection is carried out on the target list information according to a set detection method, so that a corresponding abnormality detection result is obtained; inputting the abnormal sample information corresponding to the abnormal detection result into a pre-trained pre-estimated model for evaluation processing, and outputting the on-board duration corresponding to the abnormal sample information; determining a feedback adjustment strategy according to the on-board duration; and executing the processing of the abnormal list information based on the feedback adjustment strategy. By creating a detection tool for list abnormality, abnormal sample information in each list information can be detected, and then the abnormal sample information is processed through a set feedback adjustment strategy, so that the purpose of treating the abnormal list information is achieved; by the scheme, the technical effects of forming a set of real-time alarming, feedback and adjusting hotlist management mechanism and maintaining fairness and stability of the hotlist by combining machine audit with manual audit can be achieved.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the application and together with the description, serve to explain the principles of the application.
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings that are required to be used in the description of the embodiments or the prior art will be briefly described below, and it will be obvious to those skilled in the art that other drawings can be obtained from these drawings without inventive effort.
One or more embodiments are illustrated by way of example and not limitation in the figures of the accompanying drawings, in which like references indicate similar elements, and in which the figures of the drawings are not to be taken in a limiting sense, unless otherwise indicated.
FIG. 1 is a flowchart illustrating a method for processing a list exception according to an embodiment of the present disclosure;
FIG. 2 is a flowchart illustrating another method for processing a list exception according to an embodiment of the present disclosure;
FIG. 3 is a flowchart illustrating another method for processing a list exception according to an embodiment of the present disclosure;
FIG. 4 is a flowchart of a method for detecting an abnormality of a list according to an embodiment of the present disclosure;
FIG. 5 is a flowchart of a prediction model of list anomaly provided in an embodiment of the present application;
FIG. 6 is a flowchart illustrating a method for managing list exceptions according to an embodiment of the present disclosure;
FIG. 7 is a schematic structural diagram of a device for processing list exceptions according to an embodiment of the present application
Fig. 8 is a schematic structural diagram of an electronic device according to an embodiment of the present application.
Detailed Description
For the purposes of making the objects, technical solutions and advantages of the embodiments of the present application more clear, the technical solutions of the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is apparent that the described embodiments are some embodiments of the present application, but not all embodiments. All other embodiments, which can be made by one of ordinary skill in the art without undue burden from the present disclosure, are within the scope of the present disclosure.
The terms "comprising" and "having" in the embodiments of the present application are used to mean that there may be additional elements/components/etc. in addition to the listed elements/components/etc.; the terms "first" and "second" and the like are used merely as labels, and are not intended to limit the number of their objects. Furthermore, the various elements and regions in the figures are only schematically illustrated and thus the present application is not limited to the dimensions or distances illustrated in the figures.
For the purpose of facilitating an understanding of the embodiments of the present application, reference will now be made to the following description of specific embodiments, taken in conjunction with the accompanying drawings, in which the embodiments are not intended to limit the embodiments of the present application.
Fig. 1 is a flowchart of a method for processing list exceptions according to an embodiment of the present application. The method is applied to the detection and treatment process of the abnormal list information. According to the diagram provided in fig. 1, the method for processing the list exception specifically includes:
s101, acquiring target list information, and carrying out anomaly detection on the target list information according to a set detection method to obtain a corresponding anomaly detection result.
The method and the device are applied to detecting the abnormal list information and effectively treating the abnormal list information. Constructing an abnormality detection tool of the list information according to a set detection rule or detection model, firstly carrying out abnormality detection on the target list information, and detecting abnormal sample information in the target list information; the method comprises the steps of estimating the on-board duration of abnormal sample information by utilizing a pre-trained estimation model, further calculating adjustment parameters according to the on-board duration, determining feedback adjustment strategies according to the adjustment parameters, and effectively adjusting the ranking positions of the on-board lists of the abnormal sample information according to different feedback adjustment strategies, so that the technical effects of forming a set of hot-board management mechanism for real-time alarming, feedback and adjustment and maintaining fairness and stability of a hot-board by combining machine audit with manual audit are achieved.
The target list information may be understood as a specified ranking list, including sample information of each news ranked in the ranking list. The detection method is understood to be a built abnormality detection tool. The anomaly detection can be understood as detection and analysis of the same position in the target list information in terms of list duration, list position variation range, popularity ranking consistency, primary list and secondary list anomaly and the like. The abnormal detection result can be understood as abnormal sample information in the target list information obtained by the detection tool.
Further, firstly downloading or reading target list information to be detected, carrying out anomaly detection on the target list information according to a set detection tool, analyzing whether all sample information contained in the target list information has behaviors such as brushing, abnormal stir-frying and the like, carrying out anomaly detection on the target list information, outputting the detected anomaly sample information as an anomaly detection result, and further obtaining the anomaly sample information contained in the target list information through the set detection tool so as to prepare for adjusting the anomaly sample information in the next step.
S102, inputting the abnormal sample information corresponding to the abnormal detection result into a pre-trained estimated model for evaluation processing, and outputting the on-board duration corresponding to the abnormal sample information.
The abnormal sample information can be understood as news information of abnormal situations in the target list information. The prediction model can be understood as a model for predicting the on-board duration of each news in the list information. The on-board duration can be understood as a stay time corresponding to each piece of news information in the target list information as a position.
Further, after an abnormality detection result is obtained, the abnormality sample information is predicted by utilizing a pre-trained and built prediction model, and the on-board duration of the abnormality sample information prediction is obtained and is used as a data basis for determining an adjustment strategy in the next step.
S103, determining a feedback adjustment strategy according to the on-board duration.
The feedback regulation strategy is understood herein as a grooming or maintenance strategy for abnormal sample information.
Further, according to the difference of the on-board duration obtained by prediction of the abnormal sample information, comparing the on-board duration corresponding to the normal sample information, and determining a feedback adjustment strategy according to the difference of the position adjustment amplitude, wherein the larger the difference value of the on-board duration and the on-board duration is, the farther the sample information deviates from the normal ranking position.
S104, executing the processing of the abnormal on-board information based on the feedback adjustment strategy.
The process described herein is understood to mean a position adjustment operation beyond the normal ordering position.
Further, according to different position adjustment degrees, different feedback adjustment strategies are divided, and according to the descending order adjustment strategies, abnormal sample information is used as a ranking position to carry out descending weight processing; the abnormal sample information is taken as a ranking position to be subjected to weighting treatment according to an ascending order adjustment strategy, so that the problem that target list information is unfair due to various reasons of the abnormal sample information is solved, and therefore, the technical effects of forming a set of hot list management mechanism for real-time alarm, feedback and adjustment and maintaining fairness and stability of a hot list by combining machine audit with manual audit can be achieved.
According to the method for processing the list abnormality, the target list information is obtained, abnormality detection is carried out on the target list information according to the set detection method, and a corresponding abnormality detection result is obtained; inputting the abnormal sample information corresponding to the abnormal detection result into a pre-trained estimated model for evaluation processing, and outputting the on-board duration corresponding to the abnormal sample information; determining a feedback regulation strategy according to the on-board duration; and executing the processing of the abnormal on-board information based on the feedback adjustment strategy. By creating a detection tool for list abnormality, abnormal sample information in each list information can be detected, and then the abnormal sample information is processed through a set feedback adjustment strategy, so that the purpose of treating the abnormal list information is achieved; by the scheme, the technical effects of forming a set of real-time alarming, feedback and adjusting hotlist management mechanism and maintaining fairness and stability of the hotlist by combining machine audit with manual audit can be achieved.
FIG. 2 is a flowchart illustrating another method for processing a list exception according to an embodiment of the present disclosure. The method is applied to the detection and treatment process of the abnormal list information. Fig. 2 is presented on the basis of the above embodiment. According to the diagram provided in fig. 2, the method for processing the list exception specifically further includes:
s201, acquiring target list information, and performing anomaly detection on the target list information according to the set detection rule.
S202, acquiring target list information, and performing anomaly detection on the target list information according to a set detection model.
The method and the device are applied to detecting the abnormal list information and effectively treating the abnormal list information. Constructing an abnormality detection tool of the list information according to a set detection rule or detection model, firstly carrying out abnormality detection on the target list information, and detecting abnormal sample information in the target list information; the method comprises the steps of estimating the on-board duration of abnormal sample information by utilizing a pre-trained estimation model, further calculating adjustment parameters according to the on-board duration, determining feedback adjustment strategies according to the adjustment parameters, and effectively adjusting the ranking positions of the on-board lists of the abnormal sample information according to different feedback adjustment strategies, so that the technical effects of forming a set of hot-board management mechanism for real-time alarming, feedback and adjustment and maintaining fairness and stability of a hot-board by combining machine audit with manual audit are achieved.
The target list information may be understood as a specified ranking list, including sample information of each news ranked in the ranking list. Such as a microblog host list or an entertainment list. The detection rule is understood to be a detection item or detection content defined in the construction of the abnormality detection tool. The detection model is understood herein to be a classification detection or screening model created in the construction of an anomaly detection tool. The anomaly detection can be understood as detection and analysis of the same position in the target list information in terms of list duration, list position variation range, popularity ranking consistency, primary list and secondary list anomaly and the like.
Further, firstly downloading or reading target list information to be detected, respectively carrying out anomaly detection on the target list information according to a set detection rule or a detection model, analyzing whether all sample information contained in the target list information has behaviors such as brushing, abnormal stir-frying and the like, and carrying out anomaly detection on the target list information.
Optionally, some exemption rules, such as "sudden burst money event exemption", etc. may be set, and privileges may be set for some international political sudden hot events, so as to avoid the abnormal detection process.
S203, detecting the variation range of the target list information in the list position according to the set first detection rule.
S204, detecting the same position in-list duration of the target list information according to the set second detection rule.
S205, performing hotness ranking consistency detection on the target list information according to the set third detection rule.
S206, performing primary and secondary list abnormality detection on the target list information according to the set fourth detection rule.
Further, according to some disorder of the list of the target list information, the detection of the change range of the list position can be set, for example, the detection of more than 8 positions above and below the list position of sample information corresponding to a certain news in one minute is set as abnormal sample information; or setting the same position in a list duration detection, for example, detecting that sample information corresponding to news in the last 24 hours is fixed at the same position of a list for more than 1 hour, and judging the sample information as abnormal sample information; or the detection of the ranking consistency of the hotness can be set, and each piece of news information in the target list information is detected through the remarkable contrast of the hotness change curve and the ranking change curve; or detecting abnormality of the main and auxiliary charts, for example, detecting whether the same two hot search orders in the micro-blog main chart and the entertainment hot chart are consistent, if the two hot search orders are inconsistent, judging the corresponding news information as abnormal sample information, and performing omnibearing abnormality detection on the target chart information by setting detection rules through various angles.
S207, obtaining target characteristics corresponding to the target list information.
S208, inputting the target features into a pre-trained random forest classification model to perform anomaly detection.
Further, according to some disorder images of the target list information, multi-dimensional explicit structural features such as fermentation time before the list, time point before the list, track after the list, position fluctuation range of the list, time length after the list and the like can be set as target features, the detected abnormal negative samples are combined, the extracted target features are input into a random forest classification model, classification processing is carried out on the news information category of each day in the target list information, and abnormal detection of the target list information based on a detection model is achieved.
S209, when abnormal sample information is detected according to the detection rule, a corresponding rule abnormal result is obtained.
S210, when abnormal sample information is not detected according to the detection rule, a corresponding rule normal result is obtained.
The abnormal detection result can be understood as abnormal sample information existing in the target list information obtained by the detection tool.
Further, according to the set detection rule, carrying out abnormal detection on the target list information, judging whether the sample information contained in the target list information has abnormal ranking phenomenon, and when abnormal sample information is detected, marking the detection result as a rule abnormal result so as to prepare for the next step of adjusting the abnormal sample information; and when no abnormal sample information is detected, characterizing that no list abnormality exists in the current target list information, and marking the detection result as a regular normal result.
S211, when abnormal sample information is detected according to the detection model, a corresponding classified abnormal result is obtained.
S212, when abnormal sample information is not detected according to the detection model, a corresponding classification normal result is obtained.
Further, according to the set detection model, carrying out abnormal detection on the target list information, judging whether the sample information contained in the target list information has abnormal ranking phenomenon, and when abnormal sample information is detected, marking the detection result as a classification abnormal result so as to prepare for the next step of adjusting the abnormal sample information; when abnormal sample information is not detected, the fact that no list abnormality exists in the current target list information is characterized, the detection result is marked as a classification normal result, and the target list information is not processed.
S213, inputting the abnormal sample information corresponding to the abnormal detection result into a pre-trained pre-estimated model for evaluation processing, and outputting the on-board duration corresponding to the abnormal sample information.
The abnormal sample information can be understood as news information of abnormal situations in the target list information. The prediction model can be understood as a model for predicting the on-board duration of each news in the target list information. The on-board duration can be understood as a stay time corresponding to each piece of news information in the target list information as a position.
Further, after an abnormality detection result is obtained, the abnormality sample information is predicted by utilizing a pre-trained and built prediction model, and the on-board duration of the abnormality sample information prediction is obtained and is used as a data basis for determining an adjustment strategy in the next step.
S214, determining a feedback adjustment strategy according to the on-board duration.
The feedback regulation strategy is understood herein as a grooming or maintenance strategy for abnormal sample information.
Further, according to the difference of the on-board duration obtained by prediction of the abnormal sample information, the on-board duration corresponding to the normal sample information is compared, the larger the difference value of the on-board duration and the on-board duration is, the farther the sample information deviates from the normal ranking position, and the feedback adjustment strategy is determined according to the difference of the position adjustment amplitude.
S215, processing the abnormal on-board information based on the feedback adjustment strategy.
The process described herein is understood to mean a position adjustment operation beyond the normal ordering position. For example, the adjustment is made for an anomaly that is faster in the rise of the position of the leaderboard, or for an anomaly that is faster in the fall of the position of the leaderboard.
Further, according to different position adjustment degrees, different feedback adjustment strategies are divided, and abnormal sample information is used as a ranking position to be subjected to descending treatment according to descending adjustment strategies; according to the ascending order regulation strategy, the abnormal sample information is used as a ranking position to be promoted, and the problem that target list information is unfair due to various reasons of the abnormal sample information is further changed, so that the technical effects of forming a set of hot list management mechanism for real-time alarming, feedback and regulation and maintaining fairness and stability of a hot list by combining machine audit with manual audit can be achieved.
According to the method for processing the list abnormality, different detection rules and detection models are set, target list information is detected in multiple directions to obtain abnormal sample information, the abnormal sample information is input into a prediction model for predicting the list duration, the list duration of the abnormal sample information is predicted according to characteristic data, adjustment parameters are determined according to the list duration, different feedback adjustment strategies are set according to the adjustment parameters, the abnormal list information is processed based on the feedback adjustment strategies, and the abnormal sample information is taken as a ranking position to be processed according to the descending adjustment strategies; according to the ascending order regulation strategy, the abnormal sample information is used as a ranking position to be promoted, and the problem that target list information is unfair due to various reasons of the abnormal sample information is further changed, so that the technical effects of forming a set of hot list management mechanism for real-time alarming, feedback and regulation and maintaining fairness and stability of a hot list by combining machine audit with manual audit can be achieved.
FIG. 3 is a flowchart illustrating another method for processing a list exception according to an embodiment of the present disclosure. The method is applied to the detection and treatment process of the abnormal list information. Fig. 3 is presented on the basis of the first embodiment. According to the diagram provided in fig. 3, the method for processing the list exception specifically further includes:
S301, acquiring history list information and acquiring history list features corresponding to the history list information.
The history list information can be understood as training samples for constructing the detection tool and training the pre-estimated model. The history list information can be all list information obtained after accumulating for a certain time, so that the aim of training the number of samples to achieve a training model is fulfilled.
The history list feature may be understood as detection data for detecting abnormal sample information, for example, a multi-dimensional explicit structural feature such as a range of variation of a position of a list, a duration of the same position of the list, a ranking consistency of a popularity, an anomaly of a primary list and a secondary list, an avoidance of sudden burst events, a fermentation time before a list, a time point of the list, a track of the list, a range of variation of the position of the list, a duration of the list, and the like, which are included in the history list information.
Further, before abnormality detection is performed on the target list information, a detection tool and an estimation model are built by utilizing the accumulated historical list information. Firstly, the accumulated history list information is obtained, and all history list characteristics contained in the history list information are extracted to prepare for detecting the history list information in the next step.
S302, setting up a rule detection tool based on the detection rules set by the history list information and setting up a model detection tool based on the history list characteristics.
S303, carrying out anomaly detection on the history list information based on the model detection tool and the rule detection tool to obtain a history anomaly result.
And constructing a detection tool according to the set detection rules and the detection model, carrying out anomaly detection on the obtained historical list information to obtain a historical anomaly result, and providing a training negative sample and characteristic data for the next training pre-estimated model. And training a classification discrimination model based on random forests by combining the abnormal negative samples corresponding to the detected historical abnormal results. Thus, a batch of detection tools for the messy images is formed, and the negative samples of the messy images are accumulated for other models.
Optionally, some exemption rules, such as "sudden burst money event exemption", are also included, privileges are set for some international political sudden hot events, and the abnormality detection process is avoided.
In a possible effort scenario, fig. 4 is a flowchart of a method for detecting a list abnormality according to an embodiment of the present application. According to the diagram provided in fig. 4, various detection tools for the top and bottom list anomalies are formed through rule extraction and algorithm models. As shown in fig. 4, the rule-based detection method includes: detecting the variation range of the position of the list (for example, the position is more than 8 positions up and down within one minute), detecting the same position in the length of the list (for example, the information is fixed at the same position of the list for more than 1 hour within the last 24 hours), detecting the ranking consistency of the list (for example, the ranking change curve and the ranking change curve have obvious contrast), detecting the abnormality of the main list and the auxiliary list (for example, the ranking of the same two hot searches in the micro-blog list and the entertainment hot list are inconsistent), and the like, and meanwhile, some exemption rules are included, for example, exemption for sudden burst events; the detection method based on the model is to train a classification discrimination model based on random forests according to multidimensional explicit structural features such as fermentation time before the top, top time point, top track, top position fluctuation range, top duration and the like in combination with detected abnormal negative samples. Thereby forming a detection tool for the clutter and accumulating a negative sample of the clutter for use by other models.
S304, inputting the historical abnormal information corresponding to the historical abnormal result into a pre-training model for pre-training, and obtaining a trained pre-estimation model.
Based on the list information characteristics and the accumulated abnormal negative samples, a pre-estimated model is formed. Based on the features of the first time of ranking, the highest position of ranking, the fermentation time before ranking, the ranking time length and the like, regression is carried out by combining a weighted least square method, an estimated model of the ranking time length is constructed, and the estimated time of news information under the current features in ranking can be estimated.
S305, acquiring target list information, and carrying out anomaly detection on the target list information according to a set detection method to obtain a corresponding anomaly detection result.
Further, firstly downloading or reading target list information to be detected, carrying out anomaly detection on the target list information according to a trained detection tool, analyzing whether all sample information contained in the target list information has behaviors such as brushing, abnormal stir-frying and the like, carrying out anomaly detection on the target list information, outputting the detected abnormal sample information as an anomaly detection result, further obtaining the abnormal sample information contained in the target list information through a set detection tool, and preparing for adjusting the abnormal sample information in the next step.
S306, acquiring abnormal sample characteristics of the abnormal sample information.
S307, inputting the abnormal sample characteristics into a pre-trained pre-estimated model for regression processing of a weighted least square method to obtain the abnormal time length.
And extracting abnormal sample characteristics from the detected abnormal sample information by using a trained pre-estimated model, inputting the abnormal sample characteristics into the pre-trained pre-estimated model, and presuming the abnormal sample information on-board time length possibly of the abnormal on-board by using a regression processing algorithm of a weighted least square method to provide reference data for the on-board position of the abnormal sample information for the next step.
S308, corresponding normal sample information is determined according to the abnormal sample information.
S309, the abnormal sample information and the normal sample information are input into the estimated model to determine the time length gain.
S310, determining feedback adjustment parameters according to the on-board duration of the abnormal sample information and the corresponding duration gain.
For the abnormal characteristics of the abnormal negative samples, reconstructing normal sample information based on the upper limit of the characteristics, inputting the normal samples and the abnormal negative samples of the same news information together into a list duration model to estimate the stay duration gain of the abnormal phenomenon, and quantifying specific parameters as the basis of subsequent treatment.
In a possible example scenario, abnormality detection is performed through the top 50 pieces of ranking in a microblog main list, when an abnormality phenomenon exists in a piece of news is detected by using a detection tool, for example, the existence of the piece of news is detected to be in a short time, the ranking position is risen from the 50 pieces of positions to the 1 st position, the ranking time is different for each ranking position, the ranking time is different in position, the ranking time of the first position is assumed to be 3 days, the ranking time of the 50 th position is assumed to be in a normal setting time of 1 day, the news is predicted according to a prediction model to obtain a predicted time of 3 days, the difference between the ranking time of the 50 th position and the normal condition before the occurrence of abnormality is compared to obtain 2 days, namely, the abnormal rising behavior is the ranking time of the news for 2 days, the feedback adjustment weight can refer to 1/2 of the ranking time gain, and a reciprocal reference basis is provided for the next adjustment of the ranking time of news.
In one possible effort scenario, fig. 5 is a schematic flow chart of a prediction model of list anomaly provided in an embodiment of the present application. According to the diagram provided in FIG. 5, a predictive model for predicting the length of a leaderboard is formed based on the characteristics of the leaderboard information and the accumulated negative samples (detected abnormal sample information). As shown in fig. 5, based on the features of the first top position, the top highest position, the top fermentation time, the top duration, and the like, regression is performed by combining a weighted least square method, so as to construct a prediction model for predicting the top duration of the abnormal sample information, and the approximate time of the abnormal sample information under the current feature can be predicted. For the abnormal characteristics of the negative samples (detected abnormal sample information), reconstructing positive samples (sample information corresponding to abnormal sample information under normal conditions) based on the upper limit of the characteristics, inputting the positive samples and the negative samples of the same entry together into a predicted model of a list duration to predict the stay duration gain caused by the abnormality, and quantifying specific parameters to obtain feedback adjustment parameters as the basis of subsequent treatment.
And S311, determining an ascending feedback regulation strategy when the feedback regulation parameter is larger than a set parameter threshold value.
S312, determining a descending order feedback regulation strategy when the feedback regulation parameter is smaller than the set parameter threshold value.
The parameter threshold is understood herein to be a threshold that characterizes the up-and-down adjustment of the abnormal sample information at the position of the list. For example, a parameter threshold value of 1 is set, abnormal sample information with a feedback adjustment parameter of less than 1 is determined to have intentional stir-frying and brushing behaviors, and abnormal sample information with a feedback adjustment parameter of more than 1 is determined to have malicious pressing behaviors. And when the adjustment parameter is equal to the set parameter threshold, judging that the news sample is normal, and not performing any feedback adjustment processing on the in-list position of the news sample. The ascending feedback regulation strategy can be understood as regulation and control measures which are carried out under the condition that malicious pressurizing behaviors exist in abnormal sample information. The descending feedback regulation strategy can be understood as regulation treatment measures which are carried out under the condition that abnormal sample information has intentional action of brushing a list.
Further, after the prediction of the on-board duration of the abnormal sample information is performed by using the pre-estimation model, the feedback adjustment parameters are obtained by using the negative sample information of the abnormal sample information and the positive sample information which is correspondingly inferred, the size of the feedback adjustment parameters is judged, and further measures for effectively adjusting and treating the abnormal sample information are judged. And under the condition that malicious pressurizing behaviors exist in abnormal sample information, adjusting and treating at a list position by utilizing an ascending order feedback adjusting strategy. Under the condition that the abnormal sample information has intentional stir-frying and brushing actions, the descending feedback adjustment strategy is utilized to adjust and treat the abnormal sample information at the list position.
S313, based on an ascending order feedback adjustment strategy, the authority raising process is carried out on the on-board position of the abnormal sample information according to the feedback adjustment parameters.
S314, based on a descending order feedback adjustment strategy, performing weight reduction processing on the in-list position of the abnormal sample information according to the feedback adjustment parameters.
The right raising process can be understood as a process of raising and recovering the position of the abnormal sample information. The weight reduction process is understood as a process of performing a weight reduction attack on the position of the list of abnormal sample information.
Further, after the feedback adjustment mechanism is triggered, according to an ascending feedback adjustment strategy, the condition that malicious pressing behaviors exist in abnormal sample information is performed, and response ascending adjustment is performed on the position of the abnormal sample information according to adjustment parameters, so that the state that the position of the abnormal sample information is reduced due to the malicious pressing behaviors is improved, and the current position of the abnormal sample information is lifted to a position before being higher. According to a descending feedback regulation strategy, the condition that the abnormal sample information has intentional stir-frying and brushing actions is carried out, and regulation of response descending is carried out on the position of the abnormal sample information according to regulation parameters, so that the state that the original abnormal sample information suddenly rises to a very high position due to the intentional stir-frying and brushing actions is improved, and the current position of the abnormal sample information is lowered to a position close to the original normal condition or the position in the first normal condition, so that the pressing treatment on the brushing actions is achieved.
Optionally, in the ascending feedback adjustment policy process, the magnitudes of the corresponding adjustment parameter and the second parameter threshold may be analyzed to divide the adjustment degree of the abnormal sample information at the list position. For example, setting the second parameter threshold to be 2, when the adjustment parameter is 1.5, determining that the ascending feedback adjustment strategy is executed when the adjustment parameter is greater than the set parameter threshold 1, then further comparing the adjustment parameter with the second parameter threshold to determine the adjustment degree, and judging that the feedback adjustment degree is low gear adjustment because 1.5<2, and using the adjustment parameter to slightly lift the abnormal sample information at the position of the list; assuming that when the obtained adjustment parameter is 5, since 5>2, the decision feedback adjustment degree is high gear adjustment, the on-board position of the abnormal sample information is lifted to a large extent by using the adjustment parameter.
Optionally, in the descending feedback adjustment policy process, the magnitudes of the corresponding adjustment parameter and the third parameter threshold may be analyzed to divide the adjustment degree of the abnormal sample information in the list position. For example, setting the second parameter threshold to 0.5, when the adjustment parameter is 0.8, determining to execute a descending feedback adjustment strategy smaller than the set parameter threshold 1, then further comparing the adjustment parameter with the third parameter threshold to determine the adjustment degree, and judging that the feedback adjustment degree is low-gear adjustment because 0.8>0.5 is closer to the reference threshold 1, and performing small-degree reduction on the position of the abnormal sample information in the position of the list by using the adjustment parameter; if the obtained adjustment parameter is 0.2, the feedback adjustment degree is determined to be high-gear adjustment because 0.2<0.5 is farther from the reference threshold 1, and the adjustment parameter is utilized to greatly reduce the in-frame position of the abnormal sample information. And when the adjustment parameter is equal to the set parameter threshold, judging that the news sample is normal, and not performing any feedback adjustment processing on the in-list position of the news sample.
In one possible effort scenario, fig. 6 is a flowchart of a method for managing list exceptions according to the embodiments of the present application. According to the diagram provided in fig. 6, monitoring, alarming and adjusting paths for abnormal images of the upper and lower charts of the charts are constructed based on the first two abnormality detection tools and the pre-estimated model, so that a treatment mechanism of 'strong machine examination and weak person examination' is formed. As shown in fig. 6, the monitoring is performed by an anomaly detection tool to obtain a corresponding anomaly detection result, the first machine checks, determines adjustment parameters, feeds back to the manual check, automatically performs adjustment if the manual exemption rule is hit, and requires manual review if the manual exemption rule is not hit, and adjusts through post feedback. Some exemption rules can be preset through manual verification, so that manual participation is greatly reduced, and a small amount of manual review is performed on key events. The machine audit evaluates the influence of the abnormal through the pre-estimated model, marks the abnormal which generates significant influence, triggers a feedback adjustment mechanism, properly reduces or advances the weight of the abnormal sample information through adjustment parameters, adjusts and balances the position of the abnormal sample information on the list, and ensures the fairness and stability of the list.
According to the method for processing the list abnormality, provided by the embodiment of the application, the history list information is utilized to build an abnormality monitoring tool, and meanwhile, a pre-estimated model of the prediction duration is trained to obtain a trained pre-estimated model; the method comprises the steps of carrying out anomaly monitoring on target list information, inputting the monitored anomaly sample information into a pre-estimated model, predicting the on-list duration corresponding to the anomaly sample information, further obtaining the size of an adjustment parameter, determining the adjustment strategy according to the size of the adjustment parameter, carrying out weight reduction or weight lifting treatment on the on-list position of the anomaly sample information by utilizing different strategies, further adjusting and balancing the position of the anomaly sample information on the list, ensuring fairness and stability of the list, and further realizing the aim of combining machine audit with manual audit to form a set of hot list management mechanism of real-time alarm, feedback and adjustment, and maintaining the fairness and stability of the hot list.
Fig. 7 is a schematic structural diagram of a processing device for list exception according to an embodiment of the present application. The method is applied to the detection and treatment process of the abnormal list information. According to the diagram provided in fig. 7, the processing device for list exception specifically includes:
the anomaly detection module 71 is configured to obtain target list information, and perform anomaly detection on the target list information according to a set detection method to obtain a corresponding anomaly detection result;
The evaluation module 72 is configured to input abnormal sample information corresponding to the abnormal detection result into a pre-trained prediction model for evaluation, and output a ranking period corresponding to the abnormal sample information;
a determining module 73, configured to determine a feedback adjustment policy according to the on-board duration;
and a processing module 74, configured to perform processing on the abnormal on-board information based on the feedback adjustment policy.
The processing device for list exception provided in this embodiment may be the processing device for list exception shown in fig. 7, and may execute all steps of the processing method for list exception shown in fig. 1-6, thereby achieving the technical effects of the processing method for list exception shown in fig. 1-6, and detailed descriptions with reference to fig. 1-6 are omitted herein for brevity.
Fig. 8 is a schematic structural diagram of an electronic device according to an embodiment of the present application, and the electronic device 800 shown in fig. 8 includes: at least one processor 801, memory 802, at least one network interface 804, and other user interfaces 803. The various components in the electronic device 800 are coupled together by a bus system 805. It is appreciated that the bus system 805 is used to enable connected communications between these components. The bus system 805 includes a power bus, a control bus, and a status signal bus in addition to the data bus. But for clarity of illustration, the various buses are labeled as bus system 805 in fig. 8.
The user interface 803 may include, among other things, a display, a keyboard, or a pointing device (e.g., a mouse, a trackball, a touch pad, or a touch screen, etc.).
It is appreciated that the memory 802 in embodiments of the present application may be either volatile memory or nonvolatile memory, or may include both volatile and nonvolatile memory. The nonvolatile Memory may be a Read-Only Memory (ROM), a Programmable ROM (PROM), an Erasable PROM (EPROM), an Electrically Erasable EPROM (EEPROM), or a flash Memory. The volatile memory may be random access memory (Random Access Memory, RAM) which acts as an external cache. By way of example, and not limitation, many forms of RAM are available, such as Static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double Data Rate SDRAM (Double Data Rate SDRAM), enhanced SDRAM (ESDRAM), synchronous Link DRAM (SLDRAM), and Direct memory bus RAM (DRRAM). The memory 802 described herein is intended to comprise, without being limited to, these and any other suitable types of memory.
In some implementations, the memory 802 stores the following elements, executable units or data structures, or a subset thereof, or an extended set thereof: an operating system 8021 and application programs 8022.
The operating system 8021 includes various system programs, such as a framework layer, a core library layer, a driver layer, and the like, for implementing various basic services and processing hardware-based tasks. The application 8022 includes various application programs such as a Media Player (Media Player), a Browser (Browser), and the like for realizing various application services. A program for implementing the method of the embodiment of the present application may be included in the application program 8022.
In the embodiment of the present application, by calling a program or an instruction stored in the memory 802, specifically, a program or an instruction stored in the application program 8022, the processor 801 is configured to perform method steps provided by each method embodiment, for example, including:
acquiring target list information, and carrying out anomaly detection on the target list information according to a set detection method to obtain a corresponding anomaly detection result; inputting the abnormal sample information corresponding to the abnormal detection result into a pre-trained estimated model for evaluation processing, and outputting the on-board duration corresponding to the abnormal sample information; determining a feedback regulation strategy according to the on-board duration; and executing the processing of the abnormal on-board information based on the feedback adjustment strategy.
The method disclosed in the embodiments of the present application may be applied to the processor 801 or implemented by the processor 801. The processor 801 may be an integrated circuit chip with signal processing capabilities. In implementation, the steps of the above method may be performed by integrated logic circuitry in hardware in the processor 801 or by instructions in software. The processor 801 described above may be a general purpose processor, digital signal processor (Digital Signal Processor, DSP), application specific integrated circuit (Application Specific Integrated Circuit, ASIC), off-the-shelf programmable gate array (Field Programmable Gate Array, FPGA) or other programmable logic device, discrete gate or transistor logic device, discrete hardware components. The disclosed methods, steps, and logic blocks in the embodiments of the present application may be implemented or performed. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like. The steps of the method disclosed in connection with the embodiments of the present application may be embodied directly in hardware, in a decoded processor, or in a combination of hardware and software elements in a decoded processor. The software elements may be located in a random access memory, flash memory, read-only memory, programmable read-only memory or electrically erasable programmable memory, registers, etc. as well known in the art. The storage medium is located in a memory 802, and the processor 801 reads information in the memory 802 and, in combination with its hardware, performs the steps of the above method.
It is to be understood that the embodiments described herein may be implemented in hardware, software, firmware, middleware, microcode, or a combination thereof. For a hardware implementation, the processing units may be implemented within one or more application specific integrated circuits (Application Specific Integrated Circuits, ASIC), digital signal processors (Digital Signal Processing, DSP), digital signal processing devices (dspev, DSPD), programmable logic devices (Programmable Logic Device, PLD), field programmable gate arrays (Field-Programmable Gate Array, FPGA), general purpose processors, controllers, micro-controllers, microprocessors, other electronic units designed to perform the functions described herein, or a combination thereof.
For a software implementation, the techniques described herein may be implemented by means of units that perform the functions described herein. The software codes may be stored in a memory and executed by a processor. The memory may be implemented within the processor or external to the processor.
The electronic device provided in this embodiment may be an electronic device as shown in fig. 8, and may perform all steps of the method for processing a list exception as shown in fig. 1-6, so as to achieve the technical effects of the method for processing a list exception as shown in fig. 1-6, and refer to the related descriptions in fig. 1-6, which are not repeated herein for brevity.
The embodiment of the application also provides a storage medium (computer readable storage medium). The storage medium here stores one or more programs. Wherein the storage medium may comprise volatile memory, such as random access memory; the memory may also include non-volatile memory, such as read-only memory, flash memory, hard disk, or solid state disk; the memory may also comprise a combination of the above types of memories.
When one or more programs in the storage medium may be executed by one or more processors, the method for processing the list exception performed on the device side for processing the list exception is implemented.
The processor is used for executing a processing program of the list abnormality stored in the memory so as to realize the following steps of the processing method of the list abnormality executed on the processing equipment side of the list abnormality:
acquiring target list information, and carrying out anomaly detection on the target list information according to a set detection method to obtain a corresponding anomaly detection result; inputting the abnormal sample information corresponding to the abnormal detection result into a pre-trained estimated model for evaluation processing, and outputting the on-board duration corresponding to the abnormal sample information; determining a feedback regulation strategy according to the on-board duration; and executing the processing of the abnormal on-board information based on the feedback adjustment strategy.
Those of skill would further appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or combinations of both, and that the various illustrative elements and steps are described above generally in terms of function in order to clearly illustrate the interchangeability of hardware and software. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present application.
The steps of a method or algorithm described in connection with the embodiments disclosed herein may be embodied in hardware, in a software module executed by a processor, or in a combination of the two. The software modules may be disposed in Random Access Memory (RAM), memory, read Only Memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.
The foregoing embodiments have been provided for the purpose of illustrating the general principles of the present application, and are not meant to limit the scope of the invention, but to limit the scope of the invention.
Claims (8)
1. The method for processing the list abnormality is characterized by comprising the following steps:
acquiring target list information, and performing anomaly detection on the target list information according to a set detection method to obtain a corresponding anomaly detection result;
inputting the abnormal sample information corresponding to the abnormal detection result into a pre-trained pre-estimated model for evaluation processing, and outputting the on-board duration corresponding to the abnormal sample information;
determining a feedback adjustment strategy according to the on-board duration;
executing the processing of the abnormal in-process information based on the feedback adjustment strategy;
acquiring abnormal sample characteristics of the abnormal sample information;
inputting the abnormal sample characteristics into a pre-trained pre-estimated model to carry out regression processing of a weighted least square method, so as to obtain abnormal list duration;
determining corresponding normal sample information according to the abnormal sample information;
inputting the abnormal sample information and the normal sample information into the pre-estimated model to determine a time length gain;
determining a feedback adjustment parameter according to the on-board duration of the abnormal sample information and the corresponding duration gain;
when the feedback regulation parameter is larger than a set parameter threshold value, determining an ascending order feedback regulation strategy;
And determining a descending order feedback regulation strategy when the feedback regulation parameter is smaller than a set parameter threshold value.
2. The method recited in claim 1, wherein the anomaly detection of the target list information according to a set detection method includes:
performing anomaly detection on the target list information according to a set detection rule;
and/or the number of the groups of groups,
and carrying out anomaly detection on the target list information according to a set detection model.
3. The method recited in claim 2, wherein the anomaly detection of the target list information according to the set detection rules includes:
detecting the variation range of the target list information in the list position according to a set first detection rule;
and, a step of, in the first embodiment,
detecting the same position on-board duration of the target list information according to a set second detection rule;
and, a step of, in the first embodiment,
performing hotness ranking consistency detection on the target list information according to a set third detection rule;
and, a step of, in the first embodiment,
and detecting the abnormality of the main list and the auxiliary list of the target list according to the set fourth detection rule.
4. The method recited in claim 2, wherein the anomaly detection of the target list information according to a set detection model includes:
Acquiring target characteristics corresponding to the target list information;
and inputting the target characteristics into a pre-trained random forest classification model to perform anomaly detection.
5. The method according to claim 3 or 4, wherein the obtaining the corresponding abnormality detection result includes:
when abnormal sample information is detected according to the detection rule, a corresponding rule abnormal result is obtained;
when abnormal sample information is not detected according to the detection rule, a corresponding rule normal result is obtained;
and/or the number of the groups of groups,
when abnormal sample information is detected according to the detection model, a corresponding classified abnormal result is obtained;
and when abnormal sample information is not detected according to the detection model, a corresponding classification normal result is obtained.
6. The method of claim 1, wherein the performing processing of the abnormal sample information based on the feedback adjustment strategy comprises:
based on an ascending order feedback adjustment strategy, carrying out weighting processing on the on-board position of the abnormal sample information according to the feedback adjustment parameters;
and based on the descending feedback adjustment strategy, performing weight reduction processing on the in-list position of the abnormal sample information according to the feedback adjustment parameters.
7. The method of claim 1, wherein prior to obtaining the target feature corresponding to the target list information, the method further comprises:
acquiring history list information and acquiring history list characteristics corresponding to the history list information;
setting up a rule detection tool based on the detection rule set by the history list information, and setting up a model detection tool based on the history list characteristics;
performing anomaly detection on the history list information based on the model detection tool and the rule detection tool to obtain a history anomaly result;
and inputting the historical abnormal information corresponding to the historical abnormal result into a pre-training model for pre-training to obtain a trained pre-estimation model.
8. The utility model provides a list exception handling device which characterized in that includes:
the anomaly detection module is used for acquiring target list information, and carrying out anomaly detection on the target list information according to a set detection method to obtain a corresponding anomaly detection result;
the evaluation module is used for inputting the abnormal sample information corresponding to the abnormal detection result into a pre-trained estimated model for evaluation processing and outputting the on-board duration corresponding to the abnormal sample information;
The determining module is used for determining a feedback adjustment strategy according to the on-board duration;
the processing module is used for executing the processing of the abnormal on-board information based on the feedback adjustment strategy;
the evaluation module is further used for acquiring abnormal sample characteristics of the abnormal sample information;
inputting the abnormal sample characteristics into a pre-trained pre-estimated model to carry out regression processing of a weighted least square method, so as to obtain abnormal list duration;
the determining module is further configured to determine corresponding normal sample information according to the abnormal sample information;
inputting the abnormal sample information and the normal sample information into the pre-estimated model to determine a time length gain; determining a feedback adjustment parameter according to the on-board duration of the abnormal sample information and the corresponding duration gain; when the feedback regulation parameter is larger than a set parameter threshold value, determining an ascending order feedback regulation strategy; and determining a descending order feedback regulation strategy when the feedback regulation parameter is smaller than a set parameter threshold value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310853781.8A CN116578942B (en) | 2023-07-12 | 2023-07-12 | Method and device for processing list exception |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310853781.8A CN116578942B (en) | 2023-07-12 | 2023-07-12 | Method and device for processing list exception |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116578942A CN116578942A (en) | 2023-08-11 |
| CN116578942B true CN116578942B (en) | 2023-12-22 |
Family
ID=87534495
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310853781.8A Active CN116578942B (en) | 2023-07-12 | 2023-07-12 | Method and device for processing list exception |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116578942B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103559208A (en) * | 2013-10-10 | 2014-02-05 | 北京智谷睿拓技术服务有限公司 | Application ranking fraud detection method and system |
| CN107315773A (en) * | 2017-05-26 | 2017-11-03 | 浙江大学 | A kind of fast searching method of the influence factor of ranking trend and ranking stability |
| CN107391548A (en) * | 2017-04-06 | 2017-11-24 | 华东师范大学 | A kind of Mobile solution market brush list user's group detection method and its system |
| CN108495195A (en) * | 2018-03-13 | 2018-09-04 | 武汉斗鱼网络科技有限公司 | A kind of network direct broadcasting ranking list generation method, device, equipment and storage medium |
| CN114461920A (en) * | 2022-02-21 | 2022-05-10 | 富途网络科技(深圳)有限公司 | Data processing method, device, medium and equipment for list information recommendation |
| CN115454698A (en) * | 2022-09-16 | 2022-12-09 | 唯品会(广州)软件有限公司 | Commodity ranking list checking method and device, storage medium and computer equipment |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103761228B (en) * | 2013-10-10 | 2017-10-13 | 北京智谷睿拓技术服务有限公司 | The rank threshold of application program determines that method and rank threshold determine system |
-
2023
- 2023-07-12 CN CN202310853781.8A patent/CN116578942B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103559208A (en) * | 2013-10-10 | 2014-02-05 | 北京智谷睿拓技术服务有限公司 | Application ranking fraud detection method and system |
| CN107391548A (en) * | 2017-04-06 | 2017-11-24 | 华东师范大学 | A kind of Mobile solution market brush list user's group detection method and its system |
| CN107315773A (en) * | 2017-05-26 | 2017-11-03 | 浙江大学 | A kind of fast searching method of the influence factor of ranking trend and ranking stability |
| CN108495195A (en) * | 2018-03-13 | 2018-09-04 | 武汉斗鱼网络科技有限公司 | A kind of network direct broadcasting ranking list generation method, device, equipment and storage medium |
| CN114461920A (en) * | 2022-02-21 | 2022-05-10 | 富途网络科技(深圳)有限公司 | Data processing method, device, medium and equipment for list information recommendation |
| CN115454698A (en) * | 2022-09-16 | 2022-12-09 | 唯品会(广州)软件有限公司 | Commodity ranking list checking method and device, storage medium and computer equipment |
Non-Patent Citations (1)
| Title |
|---|
| 基于提升系数的微博异常排名检测方法;印桂生等;哈尔滨工程大学学报(04);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116578942A (en) | 2023-08-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Žliobaitė | Measuring discrimination in algorithmic decision making | |
| US20210287119A1 (en) | Systems and methods for mitigation bias in machine learning model output | |
| WO2021139279A1 (en) | Data processing method and apparatus based on classification model, and electronic device and medium | |
| CN111125429B (en) | Video pushing method, device and computer readable storage medium | |
| CN110909984A (en) | Business data processing model training method, business data processing method and device | |
| KR20200075120A (en) | Business default prediction system and operation method thereof | |
| CN115662627A (en) | Early warning method for building early warning model of psychological health risks of students and application of early warning method | |
| CN111198897A (en) | Scientific research hotspot topic analysis method and device and electronic equipment | |
| CN111061948B (en) | User tag recommendation method and device, computer equipment and storage medium | |
| Kiran et al. | Dimensionality reduction procedure for bigdata in machine learning techniques | |
| CN116578942B (en) | Method and device for processing list exception | |
| CN109308660B (en) | Credit assessment scoring model evaluation method, apparatus, device and storage medium | |
| CN111178656A (en) | Credit model training method, credit scoring device and electronic equipment | |
| CN113919432A (en) | Classification model construction method, data classification method and device | |
| CN107786492A (en) | Network security situation evaluating method based on integrated classifier | |
| CN113537693A (en) | Personnel risk level obtaining method, terminal and storage device | |
| CN110704614B (en) | Information processing method and device for predicting user group type in application | |
| CN115660101A (en) | Data service providing method and device based on service node information | |
| Radovanović et al. | Making hospital readmission classifier fair–What is the cost? | |
| Malhotra et al. | A neuro-fuzzy classifier for website quality prediction | |
| CN113313582A (en) | Guest refusing and reflashing model training method and device and electronic equipment | |
| CN111401391B (en) | Data mining method, device and computer readable storage medium | |
| CN113221995A (en) | Data classification method, equipment and device based on semi-supervised deep classification algorithm | |
| CN112308294A (en) | Default probability prediction method and device | |
| Goethals et al. | Beyond Accuracy-Fairness: Stop evaluating bias mitigation methods solely on between-group metrics |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |